Lecture 2: Lab in Human Cognition
Todd M. Gureckis
Department of Psychology
New York University
1
Agenda for Today
•
What makes a good experiment? Alignments between
data and theory
•
•
•
•
Discuss reading
What makes a good experiment?
Experiment 1 - collect data/begin analysis
Assign reading for Experiment 1
2
First
What makes a good
experiment?
3
Where does life come from?
•
Since the 4th century B.C. the theory of spontaneous
generation dominated thinking about this question
•
Nile River floods -> enriches soil, etc... -> frogs appear ->
frogs must come out of the mud
4
Francesco Redi (1668)
5
John Needham (1745)
•
Boiled broth (thought to kill any microorganisms), then let it sit out
•
•
Still found evidence of micro-organisms
•
Rigorously verified and examined by his
peers and many agreed with him.
Revived the spontaneous generation
hypothesis
6
Pasteur solved it! (1859)
7
What went wrong and why?
A failure to detect confounding variables.
Was it sloppy science? Hard to say. The relevant
variables themselves were poorly understood.
8
Confounding Variables
Big challenge in cognitive science... What
information in the environment do people use and
how? Difficult to control, must proceed with
extreme caution.
9
2,4,6
10
11
Confirmation Bias
Tendencies of human (scientists included) to seek
positive examples of a theory rather than the things
that would refute it.
12
Falsification
Karl Popper (1935) - philosopher of science
“The Logic of Scientific Discovery”
13
Popper argued strongly against the latter, holding that scientific
theories are abstract in nature, and can be tested only indirectly, by
reference to their implications. He also held that scientific theory, and
human knowledge generally, is irreducibly conjectural or hypothetical,
and is generated by the creative imagination in order to solve
problems that have arisen in specific historico-cultural settings.
Logically, no number of positive outcomes at the level of
experimental testing can confirm a scientific theory, but a single
counterexample is logically decisive: it shows the theory, from which
the implication is derived, to be false. Popper's account of the logical
asymmetry between verification and falsifiability lies at the heart of
his philosophy of science. It also inspired him to take falsifiability as
his criterion of demarcation between what is and is not genuinely
scientific: a theory should be considered scientific if and only if it is
falsifiable.
-Wikipedia Karl Popper entry
14
Falsifiable theories can in principal be
ruled out by performing a critical
experiment.
15
Operational Definition
Some observable phenomena that is associated
with a hypothesis. Cognitive scientists must
propose hypothesis about how the mind works,
and also propose operational definitions (some
observable phenomena that would be predicted
from that hypothesis).
16
Correlational vs.
Experimental Studies
suppose...
You notice a decline in stork
populations and a decline in
human birth rate. BABIES must
come from storks!!
17
18
by David Ng
1 SELECT STORKS AND BABIES
2 EXPOSE ONLY STORKS TO
POLLUTION
3 DOES CORRELATION STILL
HOLD?
19
by David Ng
Correlational Study
Tracks the statistical relationship between two
variables, with limited ability to assess
CAUSATION. However, important tool.... suggests
hypotheses for experimental studies! Correlations
are informative also!
20
Examples
•
Individual Differences in W.M. - We think that subjects
who have higher working memory ability will perform better
on a learning task. We measure each subject’s w.m.
capacity prior to the experiment then give them the test.
We find that the higher the score on the w.m. test, the
better they are at the second test.
•
Brain imaging - BOLD signal recorded in visual cortex in
response to a visual memory task. Does this mean visual
cortex is the locus of memory? No! Correlated signal... the
memories could be elsewhere in the brain
21
Experimental Study
Directly manipulate one variable and finding the
consequence on another variable.
Independent variable - the variable the
experimenter changes (outside of subject’s control)
Dependent variable - the variable the experimenter
is measuring in response to the change.
22
Experimental Study
Experimental group - one group for which the
independent variable is specifically changes
Control group - no change or baseline change for
the independent variable. This was changes in the
dependent measure can be specifically identified
with the manipulation and not some other factor.
23
Experimental Study
Within subject - when the subject is in both the
control and experimental condition (higher power
since you control for individual difference in ability)
Between subject - different participants/subject in
the control and experimental conditions (lower
power, but removes confounds of repeated
testing).
24
Examples
•
Transcranial Magnetic
Stimulation (TMS) - We
believe some region of frontal
cortex is critical for certain
types of behavior. Using TMS
selectively “deactivate” this
region and measure behavior in
the task compared to a control.
•
Category Learning - We think
certain types of categories are
easier to learn than others.
Given each set of subjects a
different structure and measure
error rate.
25
Alignment between theory
and data.
26
Level of Analysis.
28
Marr’s Three Levels
✤
✤
✤
Computational Level
What is the goal of the process and why is it
appropriate/adapted to the environment
Representational Level
How can the computational level be realized?
How is information represented?What
algorithms/processes are used?
Implementation Level
How is it accomplished in neural circuits?
29
What’s a good research
question?
30
31
?
32
?
33
Scientific Methods in Six Easy Steps
1
Think of a reason why.
2
Figure out a way to check your reason.
3
and??
(write
it
up)
4
Now, everyone gets to dump on you.
5
Repeat until consensus formed.
6
See something.
inspired by David Ng
34
Next,
Let’s talk about this final
project (again)!
35
Broad Topics
✤
Perception
✤
Categorization (how do people learn new categories? how do people use
categories to reason about the world?)
✤
Memory (how does structure of information influence memory
performance? emotional/situational factors on memory)
✤
Learning (what factors improve human learning? effect of information
order on learning, learning task, induction mode, etc...)
✤
Decision making (how does mental fatigue influence decision making?
who do people make decision in complex task with many choice options)
✤
Language (how do people learn language-like materials? What influence
does language processes have on basic perception processes, Whorfianhypothesis, etc...)
36
Example final project’s from past years
✤
The factors influencing false memory.
✤
Influence of schemas on memory (is memory better for schema
consistent versus inconsistent information?)
✤
Sperling’s partial report paradigm (visual short-term memory)
✤
Effects of categorization on memory (does categorizing something make
you worse at remembering it?)
✤
Effects of presentation format on math problem solving (how does the
spacing between elements influence mathematical problem solving?)
✤
Sequence learning (how do people learn different types of sequence
structures?) What effect does the modality of presentation (auditory/
visual) have on sequence learning?
✤
Serial position effects in recall from memory
✤
Perception of human faces and the extraction of “prototypes” from a set
of faces
✤
Influence of active/passive learning on the acquisition of new materials.
37
Finally,
Let’s do an experiment!
38
For next time....
•
Come prepared to start analyzing Exp. 1! Boot camp has begun.
39